If there is no visual distinction between dialogue and normal text, it will be impossible to automate this. After all, the algorithms need something to work with. If you can make the distinction without reading it, it should be possible. It will need to be proofread though.
|